Managing pip in an RPM environment

darksky picture darksky · Jan 17, 2014 · Viewed 7.4k times · Source

For our prod apps, we manage all packages in RPM. In a perfect yum-only world, this is fine.

However, each language or framework recently are deploying their own package managers. For python for example, I am almost always using pip and then having to build the packages for rpm again when going to production. Same goes for npm or gems.

Is this an issue that a lot of people here are running into? What is best way to avoid all this? Are people dealing with this by just sucking it up and packaging everything manually, or is there some automated way to integrate them?

I would love to hear thoughts on this.

Answer

jhutar picture jhutar · Mar 21, 2014

I have used pyp2rpm to build rpm package for rstr module and I dislike random installations of some files via pip as well:

pyp2rpm -n rstr > ~/rpmbuild/SPECS/rstr.spec
rpmbuild -ba ~/rpmbuild/SPECS/rstr.spec

but I do not know if this is good enough for your production use.